Medical imaging is used in many applications to determine features of the composition of tissue not visible to the naked eye. The images can be displayed to a user, where the intensity or color of the image is a function of some parameter of the tissue composition. For example, computed tomography (CT) displays to the user the absorption of X-rays in the body and ultrasound displays the echo pattern produced in response to a pulsed sound wave. Of particular interest are the mechanical properties of tissue which can also be depicted in images. Changes in the mechanical properties of certain tissues can be an indication of disease. Traditional diagnostic methods have relied on the use of manual palpation to discriminate between healthy tissue and diseased regions because imaging methods were unavailable for detecting changes in mechanical properties. For example, the palpation of stiffer tissue is often the first step in the diagnosis of breast cancer and prostate cancer. A change in the mechanical properties of tissue can also be an indicator of the success or failure of therapy.
Elastography is a medical imaging technique that aims to depict elasticity, a mechanical property of tissue. Elasticity is also referred to as stiffness, or the inverse of compliance. Advanced elastography techniques can also measure the viscoelastic properties of tissue, such as viscosity and relaxation time. For this imaging technique, a mechanical excitation is applied in the proximity of the tissue of interest, such as prostate, breast, liver or any other soft organ in the body, and the resulting deformation is measured. The resulting deformation is measured with ultrasound (the method known as ultrasound elastography or USE) or Magnetic Resonance Imaging (the method known as magnetic resonance elastography or MRE). The deformation is post-processed to extract information such as viscoelastic properties (e.g., shear modulus and viscosity). The deformation or tissue strain, or alternatively, the intrinsic mechanical properties of tissue are then displayed as a map of stiffness (or other meaningful mechanical properties) of the imaged object.
Clinical uses of elastography were first demonstrated in the field of ultrasound as described in U.S. Pat. No. 5,107,837 by Ophir et. al. titled “Method and Apparatus for Measurement and Imaging of Tissue Compressibility and Compliance.” Shortly afterwards elastography was introduced in the field of magnetic resonance imaging (MRI) by Ehman and Muthupillai as described in U.S. Pat. No. 5,825,186 titled “Method for Producing Stiffness-Weighted MR Images” and U.S. Pat. No. 5,977,770 by Ehman titled “MR Imaging of Synchronous Spin Motion and Strain Waves.” In the following years elastography was shown to be of clinical value for the detection and staging of hepatic (liver) fibrosis by Sinkus et. al. “Liver fibrosis: non-invasive assessment with MR elastography” in the Journal NMR in Biomedicine 2006, pages 173-179, and Ehman et. al. “Assessment of Hepatic Fibrosis With Magnetic Resonance Elastography” in the Journal of Clinical Gastroenterology and Hepatology, volume 5, Issue 10, October 2007, pages 1207-1213. Elastography imaging of the breast has been successfully demonstrated and published by Sinkus et. al. in “Viscoelastic shear properties of in vivo breast lesions measured by MR elastography” in the Journal of Magnetic Resonance Imaging volume 23, 2005, pages 159-165. Elastography of the brain was also published by Papazoglou and Braun et. al. in “Three-dimensional analysis of shear wave propagation observed by in vivo magnetic resonance elastography of the brain” in Acta Biomaterialia, volume 3, 2007, pages 127-137. More recently, elastography of the lung was demonstrated by Ehman et. al. In U.S. Pat. No. 2006/0264736 titled “Imaging Elastic Properties of the Lung with Magnetic Resonance Elastography”. MRE of the prostate ex-vivo was demonstrated first by Dresner, Rossman and Ehman, published in the Proceedings of the International Society for Magnetic Resonance in Medicine titled “MR Elastography of the Prostate” in 1999. MRE of the prostate in-vivo was demonstrated by Sinkus et. al. and published in “In-Vivo Prostate Elastography”, Proceedings of International Society of Magnetic Resonance in Medicine, volume 11, page 586, 2003. Prostate elastography was described in U.S. Pat. No. 5,952,828, 2010/0005892, 7,034,534 and the publications referred to above and also by Kemper, Sinkus et. al. “MR Elastography of the Prostate: Initial In-vivo Application.” published in Fortschritte auf dem Gebiete der Röntgenstrahlen and der Nuklearmedizin (Advances in the area of X-ray and Nuclear Medicine), volume 176, pages 1094-1099, 2004. An alternative approach to prostate elastography used excitation applied through the rectum or the urethra as described in U.S Pat. No. and 2009/0209847, 2010/0045289 and the following publication Plewes et. al. “In Vivo MR Elastography of the Prostate Gland Using a Transurethral Actuator” Magnetic Resonance in Medicine, volume 62, 2009, pages 665-671. Alternatively, the mechanical excitation can be applied by a needle that penetrates the skin as described in U.S. Pat. No. 2008/0255444.
Quantitative elastography is an advanced elastography technique that solves an inverse problem: calculating the stiffness maps in a region of interest given excitation of the tissue and measurement of resulting motion in that region. Inverse problems are better solved over a 3D (volumetric) than a 2D (cross-sectional planar) region of interest, for example with 3D MRI and 3D ultrasound. This is because knowledge of the tissue motion over a 3D region of interest allows a more complete tissue motion model (e.g. 3D wave equation) to be used. Waves can propagate in arbitrary directions so an inverse algorithm should have 3D data in order to properly compute the wave speed, or spatial wavelength, from which the shear modulus is derived.
It should be clear that the mechanical waves induced by external exciters in most of the previous mentioned techniques vary in both space and time. An ideal measurement system would measure all three components (x,y,z) of the displacements instantaneously over a volume of interest, such that a 3D vector field of 3D displacements can be obtained at many instances in time. Such measurements would form a mathematically complete representation of the wave propagation. However, such ideal measurements systems are currently infeasible, so most previous mentioned works exploit the steady state nature of the wave propagation to build up a representation through multiple measurements over several periods of the waves. This is achieved usually by synchronizing acquisition with the exciter that is creating the waves and assuming perfect periodicity in the excitations. MR imaging is a relatively slow imaging modality, so MR elastography typically requires many minutes of acquisition time. The main advantage is that MR elastography creates high quality quantitative images of the mechanical properties of tissue that are considered the gold standard in the field of elastography. Ultrasound holds promise for faster acquisition yet it poses other challenges to overcome due to the pulse-echo nature of data acquisition and need for multiple pulses, which introduce time delays from both time of flight of the pulses and the delays between pulses. More challenges arise from the desire to acquire data over a 3D volume of interest when using a conventional ultrasound transducer that acquires data from a single 2D cross-sectional plane for a given position of the transducer. It is possible to move a conventional ultrasound transducer over a volume of interest in a freehand fashion, but the set of pulse-echo data will not in general be at equally spaced spatial and temporal locations. This makes it more challenging to use conventional inversion methods to calculate the mechanical properties of tissue from the acquired measurements. There exist methods to interpolate irregularly spaced pulse-echo data of stationary tissue into a regularly spaced volume (see Rohling et al. “Comparison of freehand three-dimensional ultrasound reconstruction techniques” in Medical Image Analysis, 1999), but there are no previous reports of also accommodating the time delays for each pulse-echo step when measuring the displacements of moving tissue. It would be beneficial to invent a method that produces the high quality results of MR elastography with a freehand motion of a conventional ultrasound transducer over a volume of interest, despite the irregular spacing of ultrasound compared to MRI.
Tissue motion, as captured by an ultrasound transducer, usually represents the motion in the axial direction with respect to the ultrasound transducer. The axial direction is defined as the direction of the sound pulse created by the transducer array. The lateral direction is defined in a 2D cross-sectional plane as along the direction of the transducer array, while the elevational direction is defined as orthogonal to the 2D cross-sectional plane. The resolution of an ultrasound image is generally highest in the axial direction and lowest in the elevational direction, so tissue motion in the axial direction is measured with the highest accuracy.
In U.S. Pat. Application No 2012/000779, by A. Baghani et al., “Elastography using ultrasound imaging of a thin volume”, the entirety of which is hereby incorporated by reference, a method is presented to acquire volumetric quantitative elastography images using either matrix arrays that can electronically steer a planar beam to form a 3D volume, such as the xMATRIX ultrasound transducer (Philips Healthcare, Andover, Mass.), or using mechanically swept linear ultrasound imaging transducers, such as the 4DL14-5/38 Linear 4D ultrasound transducer (Ultrasonix Medical Corporation, Richmond, BC), that move the imaging plane in the elevational direction in order to acquire a volumetric image. In U.S. Pat. Application No 2012/000779, the sweeping motion of the mechanically swept ultrasound transducers is synchronized with the known frequency of the tissue motion in order to generate a set of tissue displacement estimates that are regularly spaced in time and space. As described in U.S. Pat. Application No. 2012/000779, these displacement estimates can be used to compute elasticity images using techniques known in the art, such as the local spatial frequency estimator.
However, the majority of transducers used with commercial ultrasound machines create only 2D images, so it would be beneficial to extend the benefits of solutions of 3D inverse problems to machines that acquire 2D images.
A standard 2D ultrasound machine can be used to acquire 3D measurements by adding a position tracker to the transducer and then moving the transducer over a 3D region or volume of interest while acquiring ultrasound images—each tagged with one or more position tracker measurements. If the position tracker provides both position and orientation, we will call the joint set of positions and orientations a “location”, as commonly done in the robotics literature. A minimum of six numbers is needed to specify the location of an object in space. In this way, the ultrasound image data is acquired at different temporal and spatial locations. This set of ultrasound data is 3D and can be conceptualized by the analogy to stacking a deck of cards, where each card is a 2D ultrasound image. However, since the ultrasound transducer is moved by hand, the set of images will lie on irregularly spaced parallel cross-sections. What is needed is a novel method and system to use the set of tracked ultrasound data in an inverse technique to obtain quantitative elastography.
Previous work such as that performed by Lindop et al (“3D elastography using freehand ultrasound,” Ultrasound in Medicine and Biology, 32(4), 2006) demonstrates the ability to create 3D strain volumes using a freehand scanning technique. Unlike the method described in this invention that solves an inverse problem to produce quantitative elastography images, Lindop et al, have created a method using only axial strain imaging. A 2D ultrasound transducer is tracked in 3D space using an active optical tracker. As the transducer is moved, a series of cross-sectional planes are acquired. Each plane is spaced approximately 0.1 mm from its neighbors. They assume that the de-correlation between cross-sectional planes is small and that the small changes in pressure due to the user's motion will cause enough strain to create an elastogram. The strains are calculated using cross-correlation techniques to track the tissue deformation in the axial direction using the radio frequency ultrasound data. However, the Lindop et al. method does not provide the quantitative mechanical properties of tissue, but instead, provides the relative strain. The image quality provided with that method can also be compromised by the accumulation of de-correlation due to involuntary motion of the user in all five degrees of freedom that violate the assumptions of only axial compression of tissue. Considerable correction techniques should be applied after the original cross-correlation in order to create an interpretable image. It is also only capable of measuring the static mechanical properties, not the dynamic properties that require measurements over time. Such dynamic techniques are typically based on observing wave motion in tissue.
It is often the case that the frequency of tissue motion exceeds the imaging speed of commercially available ultrasound machines. Vibration frequencies can range from 10 to 300 Hz, while most commercial ultrasound machines produce images at approximately 40 Hz, depending on the depth of imaging. In order to overcome this drawback, two main methods have been developed. The first method uses techniques to speed up the effective frame rate. These techniques could include sector based imaging as described by Baghani et al (“A high-frame-rate ultrasound system for the study of tissue motions”, IEEE Ultrasonics, Ferroelectrics and Frequency Control. 57(7), 1535-1547 (2010)). Another technique involves careful selection of imaging frame rates and vibration frequencies and is described by Eskandari et al (“Bandpass sampling of high frequency tissue motion”, IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, 58(7), 2011).
A key application of elastography is surgery where elastograms can guide surgery. The prevalence of minimally invasive surgery, where a surgeon could find benefit from elastogram guidance, is growing. This type of surgery involves the surgeon using long instruments through small holes in the patient's skins. The difficulty of using these instruments to complete complicated procedures led to the development of robotic laparoscopic surgery. In particular, Intuitive Surgical Inc. has commercialized the da Vinci™ Surgical system. This surgical robot gives the surgeon 6 degrees of freedom of the position and orientation of the end effecter of the tool. This system incorporates a stereo laparoscope, allowing the surgeon to view the surgical scene in 3D. Some embodiments of the invention take advantage of both the dexterity and stereo vision systems of this robot to move and track an ultrasound transducer.
The invention disclosed herein uses a steady-state or periodic excitation to create dynamic motion within a tissue. A tracked 2D ultrasound transducer is used to image the volumetric tissue displacements to create a volume of the mechanical properties of the tissue in real-time without disrupting the natural motion of scanning used by the physician. Tracking of the ultrasound transducer can be achieved, for example, by embedding a magnetic sensor, an inertial measurement unit, or one or several cameras for localization inside the transducer. Note that the scanning motion is described here as arising from the physician's hand, but it could also arise from a robot, where the robot is either moved automatically or controlled by a human operator.